Development of Innovative Frameworks for Application Analysis in Post-Peta Scale Systems

2023 Fiscal Year Final Research Report

• Project/Area Number: 20H04193
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 60090:High performance computing-related
• Research Institution: The University of Electro-Communications
• Principal Investigator: Miwa Shinobu 電気通信大学, 大学院情報理工学研究科, 准教授 (90402940)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 高性能計算 / プロファイル / トレース / 予測

Outline of Final Research Achievements

This research project was conducted from April 2020 to March 2024 and produced various notable results related to predicting the profiles and traces of parallel applications. More specifically, we developed two methods for predicting the function call and cache miss counts as profile prediction, while we developed two methods for predicting the MPI communication and memory access traces as trace prediction. In addition, we surveyed many profilers used in various platforms such as GPUs and Intel SGX to extend the proposed methods to these platforms. We performed three collaborations with four oversea researchers in this research project. The results of this research project were partially presented in an authorized international workshop and a top journal in the field of high performance computing.

Free Research Field

高性能計算

Academic Significance and Societal Importance of the Research Achievements

本研究成果の一部は著名な国際会議や英文論文誌にて発表したことから，並列アプリケーションの性能解析に大きなインパクトを与えたと言える．本研究課題によって達成した並列アプリケーションの性能解析コストの削減は，今後の並列アプリケーション開発の速度向上とアプリケーションそのものの速度向上に繋がる成果であり，計算科学分野のさらなる発展に資するものである．